mechnical properties of rpc

8/2/2019 Mechnical Properties of Rpc

1/16

Mechanical Properties of Modified Reactive Powder Concrete

By S. Collepardi, L. Coppola, R. Troli, M. Collepardi

Synopsis: Original Reactive Powder Concrete (RPC) - in form of a

superplasticized cement mixture with silica fume, steel fibers and ground fine

quartz (150-400 m) - was studied in comparison with a modified RPCwhere a

graded natural aggregate (max size 8 mm) was used to replace the fine sand

and/or part of the cementitious binder.

Original and modified RPCwere manufactured at a plastic-fluid consistency,

cast by vibration and cured at three different conditions: a) room temperature; b)

steam-curing at 90C; c) high pressure steam-curing at 160C.

The addition of the graded aggregate does not reduce the compressive strengthprovided that the quality of the cement matrix, in terms of its water-cement ratio,

is not changed. This result is in contrast with the model proposed to relate the

high compressive strength level of RPC (200 MPa) to the absence of coarse

aggregate.

Both the original and modified RPC (with the coarse aggregate addition)

perform better - in terms of higher strength and lower driying shrinkage or creep

strain - when they are steam cured rather than cured at room temperature. This

improvement was related to a more dense microstructure of the cement matrix,

particularly in theRPCspecimens steam cured at 160C.

Keywords: Creep properties; flexural strength; high-strength concrete;

shrinkage; silica fume; steels; superplasticizers; toughness.


2/16

Silvia Collepardi is a research civil engineer of Enco, Spresiano, Italy. She is

working in the field of superplasticized concrete mixtures and has published a

couple of papers in this area.

Luigi Coppola is a research civil engineer and technical director of Enco. He

has authored numerous papers on various aspects of concrete technology,

durability and mix-design.

Roberto Troli is a research civil engineer and director of the Enco laboratory.

He is author of several papers in the field of concrete technology and in

particular of chemical and mineral admixtures.

Mario Collepardi is Professor of Materials Technology and Applied Chemistry

in the Ancona University, Italy. He is author or co-author of numerous papers on

concrete technology and cement chemistry. He is also the recipient of awards for

his contributions to the fundamental knowledge of superplasticizers and their use

in concrete.

INTRODUCTION

The term Reactive Powder Concrete (RPC) has been used to describe afiber-reinforced, superplasticized, silica fume-cement mixture with very low

water-cement ratio (w/c) characterized by the presence of very fine quartz sand

(0.15-0.40 mm) instead of ordinary aggregate (1, 2). In fact, it is not a concrete

because there is no coarse aggregate in the cement mixture. The absence of

coarse aggregate was considered by the inventors to be a key-aspect for the

microstructure and the performance of the RPC (1, 2) in order to reduce the

heterogeneity between the cement matrix and the aggregate. However, due to the

use of very fine sand instead of ordinary aggregate, the cement factor of the RPC

is as high as 900-1000 kg/m3. This unusual cement factor could increase drying-

shrinkage and creep strain of the RPCwith respect to ordinary concrete with

cement factor usually in the range of 300-500 kg/m3.

The main purpose of the present investigation was to modify RPCincluding

some coarse aggregate in the mixture and to study the influence of the coarse

aggregate on the properties of cement mixtures in terms of required mixing

water, compressive and flexural strength, shrinkage, swelling and creep.

Moreover, with respect to the original manufacturing process (1) - where

additional sophisticated treatments are also used to remove the excess of mixing

water through compacting pressure of the mixture in the moulds before and

during setting, or to prolong the heating process at 90-400C to dry the hardened

material - traditional casting by vibration of fresh mixtures and conventionalcuring processes were adopted in the present investigation.


3/16

EXPERIMENTAL

Materials

A C3A-free portland cement (Blaine fineness 340m2/kg), combined with a

grey-coloured un-densified silica fume, was used as cementitious binder (Table

1). Details on chemical composition and properties of these materials were also

given in previous papers (2, 3).

Fine ground quartz sand (0.15-0.40 mm) with a specific gravity of 2.75 g/cm

3

was used for manufacturing the RPC mixture according to the original

composition given by Richard and Cheyrezy (1). For modified RPCmixtures,

well graded natural aggregate (maximum size: 8 mm, and specific gravity 2.75

g/cm3) was used to replace a part or the whole volume of the fine sand and/or a

part of the cementitious binder. Figure 1 shows the particle size distribution of

the ground fine sand in agreement with the original RPCand that of the natural

aggregate - a limestone based rock - used in the present investigation to replace

the fine sand and manufacture theRPC.

Steel fibers - 13 mm long with a diameter of 0.18 mm and an aspect ratio of

72 - were used.

An acrylic polymer (AP), in form of a 30% aqueous solution, was used as

superplasticizer. Details on the performance of this superplasticizer, with

respect to other superplasticizers, are given elsewhere (4, 5).

Concrete Mixtures

Three sets of concrete mixtures were manufactured with respect to the

originalRPCcomposition without coarse aggregate:

i) a set where the amount of ground fine quartz sand (0.15-0.40 mm) of theoriginal RPC composition was partially or totally replaced by the

natural graded aggregate (0-8 mm) without changing the cement factor

(Table 2);

ii) a set where part of the cementitious binder (cement + silica fume) wasreplaced by the graded aggregate (0-8 mm) without changing the amount

of fine sand (Table 3);

iii) a set where part of the cementitious binder and the whole of fine sandwas replaced by the graded aggregate (Table 4).

For each concrete mixture, a proper amount of mixing water - including thatof the superplasticizer aqueous solution - was used to attain to the same

workability level corresponding to a plastic-fluid consistency: 150-155 mm

according to a modified flow table test described in other papers (2, 3).


4/16

All concrete specimens were consolidated by vibration and measurements

were carried out after an adequate curing time.

Curing

Concrete specimens were moist cured in three different conditions:

a) room temperature (always at 20C);b) steam curing: at 90C after a preliminary curing at 20C for 6 hours;c) high-pressure steam curing (autoclave process) at 160C after a

preliminary curing at 20C for 24 hours.

Details of these curing process are shown in Fig. 2.

Measurements

The following properties were measured on hardened concrete specimens:

compressive strength on cube specimens (40 mm); observation of the microstructure by scanning electron microscopy; flexural strength and stress-strain curves by symmetrical two point loading

test on beam specimens (150x150x600 mm);

flexural strength by one central loading test on prism specimens(40x40x160 mm);

steel-concrete bond measured through pull-out test on deformed steel bars(diameter 20 mm) embedded for 100 mm in a 200-mm cube specimens;

shrinkage (65% R.H.) and swelling (under water) of prism specimens(50x50x250 mm);

modulus of elasticity and creep strain on cylinder specimens (diameter of60 or 120 mm with height/diameter of 2) under stress of 13.3 or 53.0 MPa.RESULTS

The data of the present work will be examined in two separate sections:

preliminary tests on compressive strength of the mixtures shown in Tables 2 to 4,

and the additional tests on some specific mixtures.


5/16

Compressive Strength

Figure 3 shows the compressive strength of the original RPCwith ground

quartz sand only (Mixture No. 1 on Table 2) versus modified RPCwhere this

fine sand was partly or totally replaced by the graded aggregate with a maximum

size of 8 mm. (Mixture No. 2 or 3 respectively). A small strength increase was

recorded when the fine quartz sand (Mixture No. 1 in Fig. 3) was replaced by

the coarser natural graded aggregate (Mixtures No. 2 and 3 in Fig. 3) due to the

small reduction in the w/c (Table 2). The 28-day compressive strength was in

the range of 160-180 MPa for the specimens cured at room temperature orsteam-cured at 90C, and in the range of 190-210 MPa for the autoclaved

specimens cured at 160C. Therefore, the compressive strength results of the

present work do not confirm the beneficial role played by the very fine sand with

respect to a coarser aggregate, as claimed by Richard and Cheyrezy (1, 2).

Figure 4 shows the effect of the graded natural aggregate, substituted for a

part of portland cement and silica fume, on the compressive strength: due to

reduction in the cementitious material, the w/c and w/(c+sf) increased a little by

using the graded aggregate (Table 3) and consequently the compressive strength

of the modified RPC(Mixture No. 5 and 6 in Fig. 4) was a little lower with

respect to the original one (Mixture No. 4 in Fig. 4).Figure 5 shows the compressive strength of the modified RPCmixtures, all

without very fine ground sand, and with reduced amounts of cement and silica

fume (Mixtures 7, 8 and 9 of Table 4) with respect to that of the original RPC

(Mixture No. 1, Table 2). Again, due to the increase in the water-cement and

water-binder ratios (Tables 2 and 4), the compressive strength decreased by

increasing the aggregate to cement ratio from 1.10 (original RPC) to

1.35-2.08 (modifiedRPC). However, the reduction in the compressive strength

of the autoclaved specimens at 160C was negligible with respect to that

recorded for the specimens cured at lower temperatures (20C or 90C).

Additional Tests

Scanning electron micrographs. In order to understand the role played by the

curing process in determining the performance of the material, the microstructure

ofRPCwas studied by SEM. To facilitate the observation of the cement matrix,

specimens without steel fibers were manufactured. Figure 6 shows typical

fracture surfaces ofRPCcured at 20C, steam-cured at 90C or autoclaved at

160C. All specimens were observed at 7 days. The microstructure of the

autoclaved specimen (Fig. 6C) was much more dense than that of the RPC

steam-cured at 90C (Fig. 6B) and this appeared a little less porous than the

material cured at 20C (Fig. 6A).


6/16

Flexural strength. Table 5 shows the influence of the graded natural

aggregate on the flexural strength with respect to the original RPC(Mixture No.

1); when all the fine ground sand was replaced by the coarser aggregate

(Mixture No. 3) there was a reduction in strength and the effect was more

significant when part of the cementitious binder and all the fine sand were

replaced by the graded aggregate (Mixture No. 8).

Flexural strength, measured on larger specimen beams (150 mm thick) test

through symmetrical two-point loading, were significantly lower than the values

measured on smaller specimens (40-mm thick) by central point loading test. This

agrees with the published data from technical literature (6). The flexural strengthdata (25-60 MPa) on RPCaccording to Richard and Cheyrezy (1) - obtained

through central point loading test and relatively small specimens (40-70 mm) -

are in good agreement with those of the present investigation of similar thickness

(40 mm).

Bending stress-strain curves. Figure 7 shows the deflection of beams under

bending stress for the original (Mixture No. 1) and modified RPC(Mixtures No.

3 and 8). A small change in the deflection curve is caused by replacing the fine

ground sand of the original RPC (Mixture No. 1) with the graded aggregate

(Mixture No. 3). A further reduction in the cementitious material (Mixture No. 8)

reduced the area under the stress-deflection curve and therefore the toughness of

the material.

Strength-bond. Table 5 shows that the steel-concrete bond, measured at 28

days on the specimens cured at 20C, was much higher in modified RPCwith

coarse aggregate than in the original one. This could be ascribed to the

interlocking between coarse aggregate and steel reinforcement which increases

the stress required to pull the bar out.

Shrinkage-swelling behaviour. Original and modified RPC mixtures, after a

7 day curing time, were exposed to a 65% R.H. air for 30 days and then

submerged under water (Fig. 8). In contrast with what happens for ordinary

concretes, there was no significant change in the shrinkage-swelling behaviour

as a function of the aggregate-cement ratio of the examined RPC. This can be

ascribed to the very dense microstructure of the cement matrix of these materials.

Moreover, both shrinkage and swelling, regardless of the mixture composition,

were lower in steam-cured and specially in autoclaved specimens with respect

to those cured at room temperature.

The shrinkage of original or modifiedRPCcured 7 days at room temperature

was about 600 10-6 after 1 month of air-exposure (65% R.H.). The change in

dimensions of the autoclaved mixture was significantly lower - about 10 times

less - than the corresponding material cured at room temperature. This is an

agreement with both the microstructural densification produced by the autoclave

curing at 160C (Fig. 6) and the published data (6). On the other hand, theshrinkage reduction of the steam-cured (90C) RPCwith respect to that of the

corresponding material cured at room temperature - less than 50% - is supported

by the SEM observation of the present paper (Fig. 6), but it is not


7/16

confirmed by other published data (6). Nevertheless, some preliminary water

absorption measurements - not shown in the present paper - indicated that steam-

curedRPCat 90C were less porous than the corresponding materials cured at

room temperature and this could justify the differences in shrinkage and swelling

as well betweenRPCcured at room temperature and that steam cured at 90C.

Modulus of elasticity and creep behaviour. Figure 9 shows creep strains

under permanent compressive stress of 13.3 or 53 MPa of the original RPC

(Mixture No. 1) and modified RPC with the graded gravel replacing all the

ground fine sand (Mixture No. 3) or substituting for part of the cementitious

binder and all the fine sand (Mixture No. 8). All the specimens were steam-cured at 90C and subjected to compressive loading at 7 days in the air

(RH=65%), i.e. at a time when the compressive strength ranged from 150 to 160

MPa.

From the immediate strain of the specimens a modulus of elasticity of about

40 GPa was calculated for all the three mixtures and this value appeared to be

lower than that recorded by Richard and Cheyrezy (56-60 GPa) for RPCwith a

compressive strength of 200 MPa (1).

The creep strain after loading of the RPCsteam-cured specimens under the

stress of 13.3 MPa (and therefore with a stress-strength ratio of about 0.09) was

much lower than that of ordinary strength concrete with compressive strength of30-40 MPa (6). The creep strain of the RPCspecimens under the stress of 53

MPa (therefore with a stress-strength ratio of 1/3 at the loading time) was of

course higher. However, the ultimate specific creep including the elastic strain

was about 35 10-6 MPa-1 regardless of the stress-strength ratio and of the

aggregate/cement ratio of the original or modifiedRPC, and then about one third

of the specific creep of ordinary strength concrete (6).

Other creep tests indicated that the specific creep of autoclaved RPC was

even lower, whereas that ofRPCcured at room temperature was of the same

order of magnitude as that of ordinary strength concretes.

CONCLUSIONS

The replacement of the fine ground quartz sand (0.15-0.40 mm) by an equal

volume of well graded natural aggregate (maximum size: 8 mm) did not change

the compressive strength of the RPC at the same water-cement ratio. These

results are not in agreement with the model proposed by Richard and Cheyrezy

since they attributed the high compressive strength level ofRPC to a better

homogeneity of the mixture in the absence of coarse aggregate.

When the graded aggregate replaced all the fine sand and part of the

cementitious binder (cement and silica fume), at a given workability level, there

was an increase in the water-cement ratio, due to the reduction in the cement

factor, and hence a corresponding decrease in the compressive strength.


8/16

Flexural strength was lower when graded coarse aggregate replaced all the

very fine sand. This effect could be explained in terms of a better homogeneity

when only very fine sand is present and then to a more effective bond-strength

between cement matrix and aggregate under shear stress produced in bending

tests.

In the presence of coarse aggregate the steel-concrete bond was increased

with respect of the original RPC and this effect could be ascribed to the

interlock developing between coarse aggregate and deformed reinforcing bar.

Steam-curing at 90C and specially high pressure steam-curing at 160C gave

a better performance ofRPC - in terms of higher strength, lower dryingshrinkage and creep strain - than the curing at room temperature. This

improvement in the performance was related with a more densified

microstructure of the cement matrix.

ACKNOWLEDGEMENT

The work in preparing the text and the Figures by Romina Boaretto and

Alessandra Galletti is acknowledged.

REFERENCES

1. Richard, P. and Cheyrezy, M.H. Reactive Powder Concretes with HighDuctility and 200-800 MPa Compressive Strength, Concrete Technology:

Past, Present, and Future, Proceedings of the V. Mohan Malhotra Symposium,

ACI SP-144, S. Francisco 1994, pp. 507-518. Editor: P.K. Mehta.

2. Coppola, L., Troli, R., Collepardi, S., Borsoi, A., Cerulli, T. and Collepardi,M. Innovative Cementitious Materials. From HPC to RPC. Part II. TheEffect of Cement and Silica Fume Type on the Compressive Strength of

Reactive Powder Concrete, LIndustria Italiana del Cemento, 707, 1996, pp.

112-125.

3. Coppola, L., Cerulli, T., Troli, R. and Collepardi, M. The Influence of RawMaterials on Performance of Reactive Powder Concrete, International

Conference on High-Performance Concrete, and Performance and Quality of

Concrete Structures, Florianopolis, 1996, pp.502-513.

4. Cerulli, T., Collepardi, M., Coppola, L., Ferrari, G., Pistolesi, C., Quek, F.and Zaffaroni, P. Zero Slump-Loss Superplasticized Concrete, 18thConference on Our World in Concrete & Structures, Singapore, 1993, pp.

73-79.


9/16

5. Collepardi, M. Superplasticizers and Air Entraining Agents: State-of-the-Art and Future Needs, Concrete Technology: Past, Present, and Future,

Proceedings of the V. Mohan Malhotra Symposium, ACI SP-144, San

Francisco 1994. Editor: P.K. Mehta, pp. 399-416.

6. Neville, A.M. Properties of Concrete, Fourth Edition, 1995, Editor:Longman Group Limited, Essex, England, p. 844.


10/16

Table 1 - Composition and properties of silica fume and portland

cement (Type V ASTM or CE I 42.5R according to ENV

197/1).

Composition (%) - Property Portland Cement Silica Fume

SiO2Al2O3Fe2O3CaO

MgOK2O

Na2O

SO3C3A

20.59

3.66

6.10

63.78

0.950.48

0.26

250

0

98.87

0.01

0.01

0.23

0.010.08

0.00

0.23

-

Blaine fineness (m2/kg) 340 -

Mean particle size (m):

without superplasticizer

with superplasticizer

-

-

13.87

0.76

Table 2 - Composition of RPC mixtures with graded aggregate (max

size = 8 mm) replacing fine sand (0.15-0.40 mm).

Mixture No. 1

(Original RPC)

2

(Modified RPC)

3

(Modified

RPC)

Portland cement (c)

Silica fume (sf)

Fine sand (fs)

Graded aggregate (ga)

Superplasticizer (dry)Steel fibers

Water (w)

934 kg/m3

234 kg/m3

1030 kg/m3

--

12.7 kg/m3187 kg/m3

215 kg/m3

933 kg/m3

234 kg/m3

539 kg/m3

489 kg/m3

12.7 kg/m3187 kg/m3

205 kg/m3

937 kg/m3

235 kg/m3

--

1031 kg/m3

12.7 kg/m3187 kg/m3

200 kg/m3

w/c

w/(c + sf)

fs/(fs + a)

a*/c

0.23

0.18

1.00

1.10

0.22

0.17

0.52

1.10

0.21

0.17

0.00

1.10

Flow table (mm) 150 150 155

* a = total aggregate =fs + ga


11/16

Table 3 - Composition of RPC mixtures with graded aggregate

replacing part of the cementitious binder (cement + silica

fume).

Mixture No. 4

(Original RPC)

5

(Modified RPC)

6

(Modified RPC)

Portland cement (c)

Silica fume (sf)

Fine sand (fs)

Graded aggregate (ga)Superplasticizer (dry)

Steel fibers

water (w)

934 kg/m3

234 kg/m3

1030 kg/m3

--12.7 kg/m3

187 kg/m3

215 kg/m3

843 kg/m3

211 kg/m3

1029 kg/m3

109 kg/m3

12.7 kg/m3

187 kg/m3

202 kg/m3

754 kg/m3

189 kg/m3

1035 kg/m3

217 kg/m3

12.8 kg/m3

188 kg/m3

204 kg/m3

w/c

w/(c + sf)

fs/(fs + a)

a*/c

0.23

0.18

1.00

1.10

0.24

0.19

0.90

1.47

0.27

0.22

0.82

1.66


* a = total aggregate =fs + ga

Table 4 - Composition of RPC mixtures with graded aggregate

replacing all fine sand and part of the cementitious material.

Mixture No. 7

(Modified RPC)

8

(Modified RPC)

9

(Modified RPC)

Portland cement (c)

Silica fume (sf)

Graded aggregate (ga)

Steel fibers

Superplasticizer (dry)

water (w)

847 kg/m3

212 kg/m3

1146 kg/m3

188 kg/m3

12.8 kg/m3

195 kg/m3

758 kg/m3

190 kg/m3

1259 kg/m3

189 kg/m3

12.9 kg/m3

192 kg/m3

665 kg/m3

166 kg/m3

1383 kg/m3

189 kg/m3

12.9 kg/m3

193 kg/m3

w/c

w/(c + sf)

a*/c

0.23

0.18

1.35

0.25

0.20

1.66

0.29

0.23

2.08


* a = total aggregate = ga


12/16


13/16


14/16


15/16


16/16

mechnical properties of rpc

Documents